Continuous Glucose Monitoring (CGM) in Prediabetes

BACKGROUND The World Health Organization (WHO) has identified diabetes as a priority non-communicable disease and launched initiatives to promote its prevention and management. Type-2 diabetes mellitus (T2DM) account for 90% of all global diabetes cases. Many countries have implemented national screening programs for early interventions and mitigate the associated complications. Such screening programs result in identifying people with pre-diabetes, a dysglycemic state in which the individuals are at increased risks of developing T2DM. Pre-diabetes include those with impaired glucose tolerance (IGT) or impaired fasting glucose (IFG) or both IGT and IFG. At least 35% of individuals with pre-diabetes will progress to T2DM within eight years without lifestyle changes. Individuals with prediabetes also have an increased risk of developing microvascular and macrovascular complications, even before progressing to frank T2DM.

The global prevalence of IGT and IFG in 2021 was 9.1% (464 million) and 5.8% ((298 million) respectively and is expected to rise to 638 million in 2045. The prevalence of prediabetes is highest in developed countries. Singapore is no exception; 14% of its adult population being identified with prediabetes.

Pre-diabetes opens a window for early intervention to de-escalate the progression to T2DM. Lifestyle interventions such as diet or exercise for adults with prediabetes can effectively mitigate their dysglycemia. A meta-analysis on pooled results from eleven randomized controlled trials reported that individuals with prediabetes who underwent lifestyle interventions had 36% lower risk of progressing to T2DM, resulting in significant gains in their quality adjusted life-years. In the Diabetes Prevention Program, the incidence of T2DM was reduced by 34% with lifestyle modifications and 18% with metformin at the 10-year follow-up. However, the potential success of effective lifestyle modifications is diluted by difficulties in adherence to healthy behaviors.

Failure to control pre-diabetes is multi-factorial. At the individual level, people with pre-diabetes lack understanding of their glycemic state and are unaware of the dietary and lifestyle measures to prevent its deterioration. They have limited access to interventions in structured programs, and do not receive adequate follow-up care or support after their diagnosis, leading to inaction. At the system level, traditional diets high in refined carbohydrates and unhealthy cooking can be difficult to modify. Social gatherings and workplace environments may also discourage healthy behaviors. The rising prevalence of obesity, sedentary lifestyle and unhealthy diets have contributed to the increasing failure to control pre-diabetes. Harnessing medical technology to enhance self-efficacy in pre-diabetes management is an option to overcome these multi-faceted barriers.

Existing tools such as continuous glucose monitoring (CGM) has been shown to improve diabetes care by providing direct and timely feedback to patients with T1DM or T2DM on their meal-related glycemic fluctuations. It is preferred over traditional capillary blood glucose monitoring due to convenience, comfort and ease of use. CGM has been deployed in prediabetes in small-scale studies in the United States of America and Canada, which reported that its short-term use in adults increased their adherence to lifestyle interventions, such as reducing food portion sizes, choosing healthier food alternatives, and increasing physical activity levels. Subsequently, the Asia-Pacific Diabetes Care Advisory Board in 2021 has recommended individuals with prediabetes to leverage on CGM to gather constant feedback on their glycemic status and to motivate their behavioral change.

Aligned with the recommendations, the Health Promotion Board in Singapore has recently launched a community-based program which aims to use CGM as one of the prediabetes interventions to improve health outcomes. However, the effectiveness and efficacy of using CGM to modify lifestyle and dietary behavior among adults with pre-diabetes, compared to routine diabetic counselling or health coaching have yet to be elucidated.

AIMS The overarching aim of the study is to assess the impact of CGM in modifying the lifestyles of adults with prediabetes. Specifically, the study primarily aims to determine the effectiveness of CGM in raising their diabetes literacy, adopting low carbohydrate and fat diet, increasing their frequency and intensities of physical activities, reducing their weight, body mass index and waist circumference over a period of 24-28 weeks.

The study also aims to

1. Explore and understand the facilitators and barriers of using CGM in prediabetes management from the key stakeholders, including both the individuals and their clinicians.
2. Determine if trial design was appropriate and was feasible with regards to i) subject recruitment and retention ii) adherence to protocol, iii) adverse events iii) patient acceptability of the intervention before planning a full scale randomized controlled trial. METHODS This study is a mixed-method study conducted at Pasir Ris Polyclinic. It is expected to take 18 months to complete. The first part of the study is a pilot two-armed, randomized controlled trial. The second part of the study involves in-depth interviews with the intervention arm participants. The third part of the study involves interviews with the healthcare providers. Due to the nature of the intervention, only the outcome assessors are blinded. Sample Size Part 1 (Pilot randomized controlled trial) The sample size calculation is based on sample size recommendation to use at least 30 subjects or greater to estimate a parameter in a pilot study. Assuming that the main trial is to be designed with 90% power and two-sided 5% significance, a pilot trial sample size per treatment arm of 25 and more will detect a small, standardized effect size that is around 0.2. The sample size will take into account possible 20% drop out, hence the final sample size is increased to 38 per arm (n=76). Part 2 and 3 (Qualitative interviews) In this qualitative part of the study, sample size is determined by data saturation, which is reached when additional data no longer reveal new themes or insights. Different guidelines exist for determining sample size for qualitative research. Bertaux (1981) proposed 15 as the smallest accepted sample. Creswell and Poth (2016) recommended 20-30 interviews to be sufficient while Morse (2000) suggested 30 as an estimated number to reach theoretical saturation. Therefore, the sample size for part 2 and 3 of the study will be a maximum of 30 patients and 30 healthcare providers respectively. Recruitment and Sampling Part 1 and 2 Recruitment of participants will be done by convenient sampling. Primary Care Physicians and Healthcare workers in Pasir Ris Polyclinic will be engaged to help identify suitable patients. Eligible patients will be verbally asked by their attending physicians to participate in the study. If the participants agree to participate, the attending doctor will complete the consultation. After doctor consultation, participants will be directed to the study team to check for eligibility, briefed on the purpose of the study and obtain written informed consent. The study procedures will be administered during the same visit. Subjects who require time to consider will be provided with the participant information sheet to read through at leisure. They will be invited to return another day if agreeable to participate and written consent will be obtained by the study team. The participants will be randomly assigned to either the intervention or control group in a 1:1 ratio according to a pre-defined computer-generated sequence using the permuted blocks randomization technique. This predefined sequence of numbers will be concealed in sequentially numbered, sealed, opaque envelopes. Envelopes are opened after eligibility has been confirmed and written informed consent obtained. Participants in the intervention arm will be asked at point of recruitment if they consent to take part in the interview. They will be briefed on the purpose of the study, their roles, their rights and that the interview will be digitally audio recorded. If agreeable, they will proceed with the interviews after de-identification. The expected participation time is 60 minutes. Part 3 The healthcare providers (HCP) caring for the participants that will be recruited. include primary care physicians, nurses and clinicians from other domains who have provided direct care to adults with prediabetes.. They will be invited to participate at the end of clinic meetings. Potential participants not present at the clinic meetings will be contacted by the study team in person. Written consent will be obtained from the healthcare providers who agree to participate in the qualitative interview. The expected participation time is 60 minutes. Participants will be reimbursed with a voucher (SGD 20) at each of the three touchpoints for the first part of the study. Participants who complete part 2 of the study will be given an additional voucher of SGD 20. Healthcare providers who agree to partake in part 3 of the study will be reimbursed with SGD 20. Part 1 (Pilot randomized controlled trial) All participants will have 3 study touchpoints at baseline (week 0), week 2-6 and week 24-28. At week 0, a healthcare professional will individually go through a prediabetes health education slide deck with participants from the control and intervention arms. The slide deck was developed by Singhealth Polyclinics in 2022, and includes topics on the pathophysiology and complications of pre-diabetes, diabetes risk assessment, and non-pharmacological management of the condition. After the health education, participants will be asked to list the food types and physical activities they feel are likely to affect glycemic control, and encouraged to make necessary lifestyle modifications. At the baseline week 0 visit, participants randomized to the intervention group will receive the Abbot Freestyle Libre system. The study team will assist the participants in applying the sensors and setting up their LibreLink app and LibreView accounts. Each participant will register for the accounts using a unique participant code as their user ID, an email address generated by the research team, and their LibreView accounts will be linked to the Principal Investigator's account. The unblinded sensor will be activated and linked to the participant's smartphone, and the participant will receive device-specific education. Participants in this group will be instructed to scan their glucose levels every 8 hourly at a minimum but can choose to scan more frequently. They will also be taught how to make diet and activity modifications to limit glucose excursions above 7.8 mmol/l, and how they could use CGM data to identify foods or physical activities more likely to affect their glucose levels, especially the ones they have listed earlier. The study team will help the participants to uninstall their LibreLink apps and Libreview accounts at the 3rd study touchpoint. Questionnaires are administered at each of the three touchpoints. These touchpoints will be conducted be face-to-face at Pasir Ris Polyclinic. The questionnaire includes the thirty-seven-item diet screener (DS) and International Physical Activity Questionnaire-Short Form for diet and physical activity measures respectively. Waist circumference, weight, and Body Mass Index (BMI) are measured at baseline and at week 24-28. Participants from the intervention and control arms will receive standard care from their physicians, and any medication titration is left to the discretion of the physicians during routine follow-up. Part 2 (In-depth interviews with the intervention arm participants) On the 3rd visit touchpoint, participants in the intervention group who have earlier consented for this part of the study will undergo in-depth interviews conducted in a quiet room at Pasir Ris Polyclinic. If the participants are unable to complete the interviews at the third study touchpoint, they may reschedule for another day, provided that all interviews are conducted within 36 weeks from the start of the intervention. The first interview will be overseen by a study team member with qualitative research experience. The principal investigator/co-investigators will conduct the interviews. The interviews are audio-recorded, and live transcription will will be transcribed verbatim using the research department's transcribing software, transcription platforms (such as transcribe.gov.sg or Note Buddy) or engagement of professional transcribers. Part 3 (Interviews/ focus group discussions with the healthcare providers) Part 3 of the study will run concurrently with part 1. Healthcare providers who cared for adults with prediabetes and consented to be interviewed will undergo either in-depth interviews or partake in a focus group discussion. This will be conducted in-person, in a quiet room at Pasir Ris Polyclinic. The first interview will be overseen by a study team member with qualitative research experience. The principal investigator/co-investigators will conduct the interviews. The interviews are audio-recorded, and live transcription will be performed with an artificial intelligence transcription tool. Research Instruments Baseline Measures The following baseline data will be collected at the 1st study touchpoint for all participants:
3. Sociodemographic data, including year of birth, gender, ethnic group, marital status, occupational status, education level and current housing type
4. Risk factors for diabetes - family history of diabetes, history of gestational diabetes, smoking
5. Co-morbidities, duration of prediabetes, metformin use (electronic medical records)
6. Measurements of body weight, body mass index (BMI), waist circumference
7. 37-item Diet Screener
8. International Physical Activity Questionnaire Short Form Sociodemographic data for HCPs The data collected from HCPs include age, gender, ethnic group, qualifications, years of practice and job title. Qualitative interview topic guides Topic guides will be used during the in-depth interviews/ focus group discussions For the intervention arm participants, it will cover four main segments: (1) general experience in using CGM; (2) challenges, constraints, potential advantages of utilising CGM; (3) the facilitators and barriers to lifestyle behavioral change with the interventions; (4) recommendations for improvement of the study procedures. For the HCPs, the following questions will be explored: (1) experience with managing adults with prediabetes; (2) knowledge and experience with CGM; (3) views on using CGM in adults with prediabetes; (4) facilitators and barriers to offering CGM as part of prediabetes management to their patients. Data Collection For quantitative data, we will use REDCAP, a database management system to capture research data electronically. It is a secure web application for building and managing databases. Participants' LibreView accounts will be linked to the Principal Investigator's account, allowing for secure and centralized data access to the glucose data for research purposes by the study team members. For qualitative data, audio recording of the interviews will be by digital voice recorder. We will be using the research department's transcribing software, transcription platforms (such as transcribe.gov.sg or Note Buddy) or engagement of professional transcribers. All identifiable information in the transcripts will be removed before analysis. Medical Device: Abbott Freestyle Libre Glucose Monitoring System and LibreLink App. FreeStyle LibreLink app ('App') when used with a FreeStyle Libre Flash Glucose Monitoring System Sensor ('Sensor') measures interstitial fluid glucose levels in people (age 4 and older). The App and Sensor are designed to replace blood glucose testing in the self-management of diabetes, including dosing of insulin but has been used in research and in diabetes prevention program in people who are normoglycemia or have prediabetes. It has been registered locally with Health Sciences Authority. The study team members will be trained in the proper administration of the device by other study team members familiar with its application in regular clinical practice. They will help the participants administer the sensors during the first study visit and dispose of the sensor applicators at the clinic. Participants will be taught how to remove it themselves after 14 days and dispose of it via standard waste collection. The device is only administered once in the study and participants are not required to re-administer it at home. Participants will register for LibreLink app and LibreView accounts using a unique participant code as their user ID and an email address generated by the research team. Participants' LibreView accounts will be linked to the Principal Investigator's account, allowing for secure and centralized data access to the glucose data for research purposes by the study team members. Data Analysis Quantitative Descriptive statistics will be generated for the baseline characteristics of the study participants, and frequencies and percentages will be presented for categorical variables while mean and standard deviation will be presented for continuous variables. Normality of the continuous variables will be assessed using the Shapiro-Wilk test, and non-normal variables will be presented in median and interquartile ranges. Comparison of body weight, BMI, waist circumference, DS scores and MET at baseline between CGM and non-CGM group will be assessed using independent t-test if the variables are normally distributed, and Mann-Whitney U test if they are skewed. General estimating equation will be performed to examine the longitudinal time points for each outcome, using Intervention as the main covariate, and baseline characteristics as other covariates. All analyses will be analyzed using IBM SPSS version 29.0 and Stata MP version 18.5. A p-value of 0.05 is considered statistically significant. Qualitative Recording will be transcribed by a digital scribe or professional transcribers. Transcription will be audited and coded by at least 2 independent study team members. Thematic analysis will be performed after familiarization with the data. After completion of open coding from the first 3 in-depth interviews using grounded theory approach, consensus on the coding framework will be reached between study team members. This framework will then be applied to the subsequent interviews and codes expanded or modified after each deliberation. In the event of discrepancies in the coding framework among the study team members, this will be resolved through consensus. The data will be coded using the NVivo-12 software.